基于时空不确定性分析的北京市农田土壤重金属镉含量等级划分

Classification of soil heavy metal cadmium content grade in Beijing farmland based on spatio-temporal uncertainty analysis

为了从时空角度探讨耕地土壤环境质量的类别划分并为差别化利用和管理土地提供可参考信息,针对多年期采样数据,采用基于时空指示克里金的方法对北京市2013年农田土壤重金属镉含量进行等级划分。其中通过自适应方法确定了等级划分最适宜的概率阈值,并基于等级错划指数估计了等级划分的不确定性。结果表明:相较于基于时空普通克里金的方法,基于时空指示克里金方法划分的等级具有更高的准确率。2013年北京市镉含量等级为超过背景值的农田区域主要分布在昌平大部分地区、平谷中部地区、大兴南部地区、房山南部等靠近城镇中心的地带,等级为低于背景值的农田区域则主要分布在延庆西部地区、怀柔北部等远离城镇中心的地带。镉含量等级的错划指数分布反映了北京市农田土壤重金属镉含量等级划分的不确定性程度。等级划分的不确定性在一定程度上受点位分布及概率阈值的影响。基于时空指示克里金的等级划分方法可为开展耕地土壤环境质量类别划分工作提供辅助支撑。

For differential use and management of land, it is important to explore the category classification of farmland soil environmental quality from a spatio-temporal perspective. In this study, based on multi-year sample data, we used the spatio-temporal Indicator Krigingmethod to classify the grade of farmland soil heavy metal cadmium content in Beijing in 2013. The self-adaptive determination method wasadopted to determine the optimal probability threshold of grade classification, and a classification error index was utilized to evaluate gradeclassification uncertainty. The results showed that the method based on the spatio-temporal Indicator Kriging had better performance withhigher accuracy than the spatio-temporal Ordinary Kriging method. The regions where the cadmium content grade was higher than thebackground value were mainly located in most areas of Changping, the central area of Pinggu, and the southern area of Daxing, which areall near the urban center. In contrast, the regions where the cadmium content grade was lower than the background value were mainly located in the west part of Yanqing and the northern part of Huairou, which are at some distance from the urban center. The classification errorindex distribution reflected the uncertainty degree of the grade classification of soil cadmium content. The uncertainty is, to an extent, affected by the distribution of sample points and probability threshold. A grade classification method based on spatio-temporal Indicator Kriging can be used to support the classification of soil environmental quality categories.